The Brady-Ryan Plan: Potential and Pitfalls

On June 24, 2016, the House Republican Tax Reform Task Force released its tax reform blueprint. The plan, which is commonly associated with House Ways and Means Committee Chair Kevin Brady, R-Texas, and House Speaker Paul D. Ryan, R-Wis., would move the tax system to a hybrid of an income tax and a Bradford X tax, a type of progressive consumption tax.¹ The partial shift to consumption taxation would significantly reduce the tax penalty on saving and investment, paving the way for increased economic growth.

Nevertheless, the Brady-Ryan plan has limitations. The tax rates on business cash flow are unnecessarily low, which reduces revenue and progressivity and creates an incentive to convert wages into business cash flow. The plan also denies expensing to inventories, thereby overtaxing them relative to other types of capital. Issues concerning the design of the plan's border adjustment must also be addressed. Whether to keep the border adjustment in the plan is a thorny question; removing the border adjustment would negate some of the plan's economic advantages, but it would preserve many of its advantages and avert some difficulties. If the border adjustment is retained, it may be desirable to relabel the business cash flow tax as a value added tax.

A. Key Provisions

The Brady-Ryan plan would combine the 10 percent and 15 percent brackets into a 12 percent bracket, merging the 28 percent bracket into the 25 percent bracket, and consolidating the top three brackets into a 33 percent bracket. The individual alternative minimum tax would be repealed. Half of interest income, dividends, and capital gains would be included in taxable income, effectively taxing those types of income at rates of 6 percent, 12.5 percent, and 16.5 percent. The blueprint does not specify whether short-term capital gains and nonqualified dividends would receive the preferential treatment.

The standard deduction would be increased to $12,000 for singles, $18,000 for heads of household, $24,000 for married couples, and would continue to be indexed for inflation. The additional standard deduction for blind and elderly taxpayers would be eliminated. The personal exemption would be replaced by a new nonrefundable $500 credit for dependents. The $1,000 child tax credit would be retained.

The itemized deductions for mortgage interest and charitable contributions would be retained. The blueprint states that the Ways and Means Committee will "evaluate options for making the current-law mortgage interest provision a more effective and efficient incentive for helping families achieve the dream of homeownership," but pledges that any changes would not affect existing mortgages. The blueprint offers a similar commitment to "develop options to ensure the tax code continues to encourage donations, while simplifying compliance and record-keeping and making the tax benefit effective and efficient." Other itemized deductions, including the deduction for state and local taxes, investment interest expense, employee business expenses, medical and dental expenses, theft and casualty losses, and gambling losses, would be eliminated.

The earned income tax credit would be retained. Tax incentives for higher education and tax-preferred savings accounts would be consolidated and simplified in an unspecified manner. The child and dependent care credit would be eliminated.

The corporate tax rate would be reduced to 20 percent and the corporate AMT would be repealed. Passthrough business income would be taxed at a maximum 25 percent rate. The research tax credit would be maintained, but the section 199 domestic production deduction and other unspecified tax preferences would be repealed.

The blueprint sets forth the general rule that expensing "will apply to both investments in tangible property (such as equipment and buildings) and intangible assets (such as intellectual property)," but identifies two exceptions. The blueprint states that expensing "will not apply to land" and "with respect to inventory, the Blueprint will preserve the last-in-first-out (LIFO) method of accounting." Businesses could deduct interest expense only against interest income, with unlimited carryforward for net interest expense.

Business taxes would be border adjusted, placing them on a destination basis. Businesses would not be allowed to deduct the costs of imported business inputs. Businesses would exclude receipts from export sales while claiming the same expense deductions that would apply for domestic sales.

The blueprint says that the estate tax would be repealed, but it does not specify whether the gift tax and the section 1014 basis step-up provision would be retained. The blueprint envisions that the taxes adopted by the ACA, including the 3.8 percent net investment income tax, would be repealed as part of a separate health policy initiative.

The blueprint offers no details about transition provisions, merely stating that "a smooth transition from the current system to the new system will be necessary" and promising to "craft clear rules to serve as an appropriate bridge from the current tax system to the new system."

B. Partial Move to Bradford X Tax

The Brady-Ryan plan's tax system can be described as a hybrid of an income tax and a modified form of the Bradford X tax. The X tax is a modification of the Hall-Rabushka flat tax, which, in turn, is a modification of the VAT.

Under a VAT, each business remits taxes based on its value added, which is its sales to other businesses and to consumers minus its purchases from other businesses. Because purchases from other businesses are immediately deducted, investment is expensed under a VAT. Because total value added in the economy is equal to consumer spending, a VAT is a consumption tax.

Robert Hall and Alvin Rabushka proposed splitting a VAT into a business cash flow tax and a household wage tax and referred to the combination of the two taxes as a flat tax. Businesses would be allowed to deduct their wage payments, so that they would pay tax on value added minus wages, which is referred to as business cash flow. A business cash flow tax is therefore equivalent to a VAT (administered using the subtraction method rather than the credit-invoice method used by nearly all VATs)² with a wage deduction. Households would be taxed on their wages. Because the combined tax base for each business and its workers would equal the value added by the business, the aggregate national tax base would remain equal to consumer spending under the Hall-Rabushka flat tax.

Aggregate consumer spending is equal to aggregate wages plus aggregate business cash flow. Future consumer spending can be financed from future wages, above-normal returns on future investments, and existing wealth. Above-normal returns, also called inframarginal returns, economic profits, and pure rents, are returns that are not available on marginal investments. Normal returns on future investments are not a source of future consumer spending because the consumer spending generated by the investments' payoffs are offset, in present discounted value, by the consumer spending sacrificed to make the investments. Under the Hall-Rabushka flat tax, the consumer spending financed by future wages is taxed by the household wage tax, and the consumer spending financed by future above-normal returns and existing wealth is taxed by the business cash flow tax.

Under the Hall-Rabushka flat tax, businesses would be taxed at a flat rate on business cash flow and households would be taxed at the same rate on wages, but only above an exemption amount. To raise the same revenue as a VAT, which does not include a wage exemption, the flat tax would require a higher tax rate. The wage exemption would make the flat tax less regressive than a VAT by lowering the tax burden on workers, particularly low-paid workers. Consumer spending financed by lower-paid workers' wages would be exempt while consumer spending financed by higher-paid workers' wages, existing wealth, and above-normal returns would be taxed.

To further promote progressivity, David Bradford proposed modifying the Hall-Rabushka flat tax to feature a full set of graduated rates on wages and possibly refundable tax credits. The top wage tax rate would equal the flat tax rate imposed on business cash flow. Bradford referred to this tax as the X tax. To raise the same revenue, the top tax rate under the X tax would need to be higher than the single tax rate under a flat tax. Compared with the flat tax, the X tax would reduce the tax burden on consumption financed by moderate-paid workers' wages while increasing the tax burden on consumption financed by high-paid workers' wages, existing wealth, and above-normal returns. Robert Carroll and I discussed the advantages of the X tax and the best way to implement it in a 2012 book.³

Although the X tax appears similar to an income tax system because it features a graduated rate tax on households and a tax on businesses, it differs from an income tax in two important ways. First, the household tax applies only to wages, not to investment income. Second, the business tax applies to real cash flows rather than to net income, with investment expensed rather than depreciated and financial flows, including interest expense, disregarded.

Those two features ensure that the X tax imposes no marginal tax burden on new investments with normal returns. The household wage tax imposes no burden on investment because capital income is not subject to the tax. The business cash flow tax also imposes no net burden on new normal return investments because the immediate tax savings from expensing offset, in present discounted value, the taxes on investment payoffs. The business cash flow tax is imposed only on existing wealth and above-normal returns.

Under the flat tax and the X tax, the burden on existing wealth would be imposed through a cash flow tax collected at the business level. Because the Federal Reserve would probably not change the consumer price level in response to the introduction of a cash flow tax,⁴ the real value of businesses' debt obligations would be unchanged, causing the entire burden to be borne by equity holders as businesses' residual claimants. For example, consider a business that holds $100 of capital and has issued $30 of debt, so that its equity is worth $70. The introduction of a 20 percent business cash flow tax without transition relief would reduce the real value of the business's capital from $100 to $80. Because the business would still owe $30 debt, the value of its equity would fall from $70 to $50. Equity holders would experience a 29 percent wealth reduction and debt holders would escape unscathed.

The Brady-Ryan plan features a graduated rate tax on wages, with a top rate of 33 percent. If that wage tax was part of an X tax, there would be a 33 percent business cash flow tax and there would be no household tax on capital income. The plan is similar to the X tax design in some ways and diverges from it in others.

The plan's business tax would resemble a business cash flow tax because most investment would be expensed and interest expense would not be deductible. However, land and inventories would not be expensed and businesses would be taxed on net interest income. The business cash flow tax rate would also be significantly lower than the top tax rate on wages. Further, the tax rate would differ across C corporations and passthrough businesses and the tax on passthrough businesses would be collected at the owner level rather than the business level. Also, households would be taxed on interest, dividends, and capital gains.

Under a pure X tax, businesses would not deduct interest expense and there would be no tax on interest income. Because the Brady-Ryan plan maintains an individual tax on interest income, its denial of a deduction for net interest expense raises additional issues. The denial of an interest deduction for corporations ensures that debt and equity are treated neutrally at the business level. Both types of finance also trigger individual-level tax, which would be imposed at preferential rates in each case. The plan therefore takes a major step toward debt-equity neutrality. However, the individual-level tax on debt holders is heavier than the corresponding tax on equity holders because equity holders can defer taxes on capital gains until realization.

The plan's denial of an interest deduction for passthrough businesses would subject debt-financed passthrough investment to a tax that equity-financed passthrough investment would not face. The denial of an interest deduction is motivated by the presence of expensing. Because of expensing, owners of passthrough investments would face a zero effective marginal tax rate on normal returns earned by passthrough businesses, making passthrough equity tax-preferred relative to holdings of debt or corporate stock. If interest was deductible, there would be a tax incentive for low-bracket individuals and tax-exempt entities to lend to high-bracket individuals, enabling the latter, who would obtain the greatest tax savings from holding passthrough equity, to hold more passthrough equity. The denial of an interest deduction would help curb the tax-motivated migration of passthrough equity to top brackets, in the same way that section 265's denial of a deduction for interest incurred to purchase tax-exempt municipal bonds attempts (with limited success) to curb the tax-motivated migration of municipal bonds to high-bracket taxpayers. Alan Auerbach of the University of California-Berkeley has provided a thorough analysis of the complicated economic issues raised by the denial of interest deductions.⁵ Some members of Ways and Means are considering modifying the plan to allow some small businesses to deduct interest expense.⁶

The plan is somewhat similar to the 2005 Advisory Panel on Tax Reform's Growth and Investment Tax plan. That plan featured a graduated rate wage tax with a 30 percent top rate, a 30 percent business cash flow tax, and a 15 percent flat rate household tax on capital income.⁷

C. Growth, Revenue, and Distribution

The blueprint describes the Brady-Ryan plan as "a 21st century tax system built for growth." It also promises a "tax system under which no income group will see an increase in its federal tax burden" and "tax reform that is revenue neutral." The blueprint states that revenue neutrality will be measured relative to a baseline in which the tax provisions expiring at the end of 2016 have been permanently extended and in which the tax increases included in the ACA have been repealed. It also states that revenue feedback resulting from macroeconomic effects will be included.⁸

It is unclear, however, if the plan is revenue neutral or if it maintains progressivity. The plan has been analyzed by the Urban-Brookings Tax Policy Center (TPC), the Tax Foundation, and my American Enterprise Institute colleague Alex Brill.⁹ The analysts made assumptions to fill in details missing in the blueprint.

The TPC estimated that, in the absence of macroeconomic feedback effects, the plan would reduce revenue by $3.1 trillion in 2016 through 2026 and $2.2 trillion during the following decade.¹⁰ Using the Penn-Wharton Budget Model, the TPC estimated that the plan would increase average GDP by 1 percent during the first decade, primarily from increases in labor supply. The TPC estimated that the plan would lower average GDP by 0.2 percent during the second decade as the larger deficits crowded out private investment. With those macroeconomic feedbacks, the TPC estimated that the plan would lose $2.5 trillion in 2016 through 2026 and $1.9 trillion during the following decade. The TPC noted that the GDP effects depend upon the openness of the economy to international capital flows. The GDP effects are more favorable if the economy is more open to international capital flows; in that case, the plan's favorable treatment of investment draws more capital to the United States and deficits crowd out less investment less because foreign capital inflows can replace the savings soaked up by deficits. Conversely, the GDP effects are less favorable if the economy is less open to international capital flows.

The Tax Foundation estimated that, in the absence of macroeconomic feedback effects, the plan would reduce revenue by $2.4 trillion in 2016 through 2025, but would have smaller revenue losses in subsequent decades. Using its Taxes and Growth Model, the foundation estimated that the plan would reduce revenue by only $200 billion in 2016 through 2025. The foundation estimated that the plan would increase the capital stock by 28 percent and would increase GDP by 9.1 percent in the long run. The estimated GDP effects are highly favorable because the Taxes and Growth Model assumes that the U.S. economy is very open to international capital flows.

Using the Open Source Policy Calculator, Brill found that the plan's individual income tax provisions reduced revenue by $227 billion from 2017 through 2026, relative to a baseline that includes the revenue from the 3.8 percent net investment income surtax. Revenue losses were declining at the end of the 10-year budget window, indicating that the plan's individual income tax provisions were close to revenue neutral in the long run. Brill's analysis did not include the corporate tax changes, the allowance of expensing for passthrough businesses, or the repeal of the estate and gift tax. His estimates assumed that taxable income and capital gains realizations are moderately sensitive to marginal tax rates.

All three analyses found that the Brady-Ryan plan would make the tax system less progressive. TPC estimated that, in 2025, after-tax income would increase by 0.1 to 0.5 percent for the bottom three quintiles, would fall 0.2 percent for the fourth quintile, and would fall 1.1 percent for the ninth decile and the bottom half of the top decile. After-tax income would increase by 1.4 percent for the 95th to the 99th percentiles, 10.6 percent for the top percentile, and 13.5 percent for the top 0.1 percent of the income distribution. The Tax Foundation found that after-tax incomes would rise 0.2 to 0.5 percent for the bottom four quintiles and would rise 1 percent for the top quintile, with a 1.5 percent increase for the top decile and a 5.3 percent for the top percentile. Brill found that the plan would give small tax cuts to the bottom four quintiles of the income distribution. Taxpayers in the top quintile would receive larger tax cuts relative to current law, although they would pay more taxes than if the high-income provisions of the 2001 and 2003 tax cuts had been made permanent. Brill noted that the income levels at which individual income tax liability becomes positive are slightly higher under the plan than under current law for married and unmarried taxpayers with no children, one child, or two children who claim the standard deduction.

The plan would reduce the fraction of taxpayers that itemized deductions from about 30 percent to about 5 percent because the standard deduction would be increased and most itemized deductions would be repealed.¹¹ Taxpayers at the same income level would be affected differently by the plan, depending on the amount of itemized deductions they claim and other factors. Tax increases would be likely for taxpayers who have mortgage interest costs and charitable deductions exceeding the plan's standard deduction amounts and who would therefore continue to itemize deductions under the new plan. Because those taxpayers would continue to itemize, they would not benefit from the increase in the standard deduction and they would lose their other itemized deductions, including the deduction for state and local taxes. They would also lose the personal exemption, with only a partial offset from the new dependent tax credit. Hudson Institute senior fellow Jeffrey Anderson has described the plan's impact on those taxpayers, presenting an example of a family of four earning $130,000 that would face a tax increase of almost $3,000.¹²

D. Selected Design Issues

The Brady-Ryan plan raises many issues that are not considered in this article. The plan could introduce financial accounting complications because it is not clear how the business cash flow tax would be classified under accounting rules. The plan would put pressure on states to reconfigure their corporate income taxes to conform to the new federal cash flow tax. Some aspects of the interaction between the business and individual tax systems remain unclear, including the determination of basis in passthrough businesses and the definition of earnings and profits used to determine the taxability of dividends. Also, the cash flow tax would not fit easily into the United States' bilateral tax treaties. I set aside those questions and many others.

1. Simplification.

The Brady-Ryan plan would offer significant simplification along many dimensions. The sharp reduction in the number of taxpayers who itemize deductions would reduce compliance and administration costs.

The blueprint claims that its tax system "will be simple enough to fit on a postcard for most Americans."¹³ The postcard displayed on page 18 of the blueprint should suffice for most taxpayers, but some taxpayers would need to file a longer form to address situations not covered by the postcard. Most strikingly, the postcard does not include a line to report income from passthrough businesses or offer any way to claim the 25 percent maximum rate that the plan would apply to such income. It also does not include lines on which to report withdrawals from tax-preferred savings accounts or gambling winnings or to claim the foreign tax credit.

The zeal to simplify the individual income tax by repealing deductions and credits resulted in a few dubious policy decisions. The repeal of the child and dependent care credit would deny tax relief for significant work-related costs.¹⁴ The repeal of the deductions for employee business expenses and moving expenses would raise similar concerns, although to a lesser extent. The plan's repeal of the tax deduction for investment interest expense¹⁵ would penalize taxpayers who borrow from one party to lend to another. The repeal of the tax deduction for gambling losses would be problematic if gambling winnings remained taxable.¹⁶

The plan would reduce the tax advantages for owner-occupied housing and charitable contributions because many fewer taxpayers would claim itemized deductions, although the taxpayers who continued to itemize would be those with the largest charitable contributions and mortgage interest costs. The reduction in marginal tax rates would also reduce the tax savings from claiming the deductions; a taxpayer in the top bracket would save only $33 for each $100 of charitable contributions and mortgage interest costs, down from $39.60 today. Economists would generally welcome the curtailment of tax breaks for owner-occupied housing. Even if there are valid reasons for the tax system to promote homeownership, the current tax preferences are ill suited for that goal. However, the reduction in incentives for charitable giving would raise significant policy concerns.

2. Negative cash flows.

A business cash flow tax imposes a zero effective marginal tax rate on a marginal new investment only if the expensing deduction provides tax savings equal in present value to the taxes imposed on the investment's future cash flows. Problems arise if businesses with negative cash flows are not accorded refundability or equally generous treatment. Without such treatment, a business that is in negative-cash-flow status when making an investment and positive-cash-flow status when it receives the investment's payoffs would face a positive effective marginal tax rate on the investment.

Along one important dimension, the plan would offer more favorable treatment for negative cash flows than a pure X tax would. Under the plan, cash flows from passthrough businesses would be flowed through to individual owners, allowing an owner with stakes in multiple businesses to net negative cash flows from one business against positive cash flows from another business. Such netting is not possible under a pure X tax, which taxes cash flows at the business level.¹⁷

In other ways, however, the plan's provisions are inadequate. The Brady-Ryan plan provides that negative cash flows will "be carried forward indefinitely and will be increased by an interest factor that compensates for inflation and a real return on capital to maintain the value of amounts that are carried forward." However, the plan undercuts those provisions by barring carrybacks and limiting the carryforward used in each year to 90 percent of that year's net cash flow.¹⁸ The cash flow tax would function more effectively if more generous treatment were provided for negative cash flows. Carroll and Viard recommended five to 10 years of carryback, as well as carryforwards with interest.¹⁹ The plan's provisions for negative cash flows would be even more inadequate for exporters, as discussed below.

E. Taxation of Financial Assets

The Brady-Ryan plan would tax businesses on their net interest income and presumably on other forms of financial income. Although a pure X tax ignores businesses' financial transactions, the plan's hybrid system cannot do so. Under the plan, individuals holding financial assets directly would pay tax, at a top rate of 16.5 percent, on the income generated by the assets. That tax could be avoided or deferred if businesses were not taxed on their income from financial assets. Financial income received by a C corporation would add to its value, but if that value increase did not result in additional dividend payments, it would escape tax until and unless the shareholders realized the resulting capital gains. Similarly, financial income received by a passthrough business would add to its value, but would not be taxed until the owner sold her stake in the business.

Financial institutions are the largest category of businesses with net financial income. Part of their financial income constitutes disguised fees for providing services for which the institutions do not charge explicit fees. The blueprint appears to recognize the need to tax those disguised fees, stating that Ways and Means "will work to develop special rules with respect to interest expense for financial services companies, such as banks, insurance, and leasing, that will take into account the role of interest income and interest expense in their business models."²⁰

Special rules will be needed because the taxation of net interest income is not a viable way of taxing the disguised fees. At first glance, it may appear to resemble the interest-spread method sometimes proposed as a way to tax disguised fees, but it differs significantly from that method.²¹

Consider a bank that has $10,000 of deposits and makes $8,000 of loans. Suppose that the risk-free rate of return is 3 percent and that the bank pays 1 percent on its deposits and earns 6 percent on its loans (net of any defaults). Because the bank pays $100 to its depositors and collects $480 from its borrowers, it has $380 net interest income. Under the interest-spread method, the bank is viewed as charging a 2 percent disguised fee to its depositors by paying them a 1 percent interest rate, which is 2 percentage points below the 3 percent market rate. Similarly, the bank is viewed as charging its borrowers a 3 percent disguised fee by collecting from them a 6 percent interest rate, which is 3 percentage points above the market rate. Depositors presumably accept below-market interest rates and borrowers presumably accept above-market interest rates because they value the services provided by the bank, such as record keeping, safe storage of deposits, and facilitation of transactions. Under the interest spread method, the bank's real cash flows are deemed to include $200 of disguised fees collected on the $10,000 of deposits and $240 of disguised fees on the $8,000 of loans, for a total of $440.

The $440 amount taxed by the interest-spread method is somewhat close to the bank's $380 net interest income. (The amount taxed by the interest-spread method would match the bank's net interest income if the amount of deposits and loans were equal; in that case, there would also be no need to specify the safe interest rate used in the calculations.) If all of the depositors and borrowers were households, the amounts might be close enough.

Severe problems arise, however, if some of the depositors or borrowers are businesses. If the bank is taxed on the disguised fees that it charges its business customers, the customers must be allowed an offsetting deduction for the disguised fees. The financial services for which those fees are paid are used to generate the business customers' own taxable output and should be deducted in the same manner as other costs of producing taxable output. Unfortunately, the plan does not allow the business depositors and borrowers to deduct the disguised fees. Because most of those businesses would presumably have net interest expenses, which would be nondeductible, they would receive no tax relief for the below-market interest rates they receive on their bank deposits or the above-market interest rates they pay on their bank loans.

The final version of the plan will therefore need to develop a separate regime for taxing financial institutions. Disguised fees for services financial institutions provide to businesses should be deducted by the business customers or should be exempt from tax.

F. Business Cash Flow Tax Rates

The statutory tax rates on business cash flow in the Brady-Ryan plan are lower than those in other proposals. The 2005 Advisory Panel's Growth and Investment Tax plan featured a 30 percent business cash flow tax rate, along with a 15 percent household tax rate on investment income. The panel also described, but did not recommend, a progressive consumption tax plan that featured a 35 percent business cash flow tax rate with no household tax on capital income.²² In our discussion of the X tax, Carroll and I assumed a 38.8 percent business cash flow tax rate, for illustrative purposes.²³

Proposed statutory tax rates have fallen as cash flow tax proposals have moved closer to the legislative process. The Nunes bill would have taxed the cash flow of C corporations at 25 percent (above the first $50,000) and would have taxed the cash flow of passthrough businesses at the owners' ordinary tax rates, with a maximum 25 percent tax rate. The Brady-Ryan plan follows the Nunes bill, except that it lowers the tax rate on C corporations' cash flow to 20 percent.

The business cash flow tax has relatively few distortions. As explained above, it is a tax on the capital in existence when the plan is implemented and on above-normal returns. Moreover, the cash flow tax falls on equity holders, an affluent group. If the cash flow tax had no border adjustment, it would discourage investments with above-normal returns from being located in the United States and would encourage income shifting. As I discuss below, however, the border adjustment removes those concerns.

To be sure, business cash flow taxes have some distortions, which increase as the tax rate rises. At higher rates, it becomes more attractive to evade tax by failing to report sales, fabricating expenses, disguising personal expenses as business expenses, and deducting hobby losses. Some of the Brady-Ryan plan's problems, including the limited tax relief for negative cash flows and the inventory tax discussed below, would become more severe at higher tax rates. Also, the disruptions from moving to the new system (for example, the disruption associated with the dollar appreciation caused by the border adjustment) would be greater at higher tax rates.

The blueprint's justification for the low cash flow tax rate conflates the cash flow tax with a corporate income tax. "This Blueprint will lower the corporate tax rate to a flat rate of 20 percent. This represents the largest corporate tax rate cut in U.S. history. . . . Economists at the OECD and elsewhere have identified the corporate income tax as the most harmful of all forms of taxation in terms of the adverse effect on growth. This dramatic reduction in the corporate tax rate will mean a dramatic reduction in the drag on American economic growth that results from the corporate income tax."²⁴ The blueprint's statement that high corporate tax rates are economically harmful is correct, but has no relevance for the appropriate tax rate for the plan's business cash flow tax.

The plan would reduce the corporate income tax rate to zero, thereby eliminating its economic drag, and replace it with a cash flow tax. The appropriate rate for the new cash flow tax depends on its economic properties, not the properties of the corporate income tax that it replaced. The blueprint states that the cash flow tax is desirable because it has few distortions. It also implies that the rate of the tax must be kept as low as possible because it has so many distortions. Logic guarantees that one of those propositions must be incorrect. Economic analysis identifies the second proposition as the culprit.

The low tax rates on business cash flow would also create significant incentives to convert wages into cash flow. Businesses that are organized as passthrough businesses under current law and in which owners perform work would have an incentive to understate wages and overstate cash flow. Also, individuals now working as employees would have an incentive to organize as passthrough businesses and become independent contractors. The blueprint says, "Sole proprietorships and passthrough businesses will pay or be treated as having paid reasonable compensation to their owner-operators. Such compensation will be deductible by the business and will be subject to tax at the graduated rates for families and individuals."²⁵ It is not clear that reasonable-compensation rules would be effective in preventing conversion of wages into business cash flow.

Because business cash flow is a more progressive, and less distortionary, tax base than wages, the tax rate on cash flow should be at least as high as the tax rate on wages. An argument could be made for taxing cash flow at a higher tax rate, although it would then be necessary to combat the conversion of cash flow into wages. An incentive to convert cash flow into wages would be a smaller problem than the opposite incentive because the numerous Americans working as employees under current law have no incentive to change their status. Instead, the smaller set of Americans who are business owners have an incentive to become employees or to report more of their cash flow as wages.²⁶

For current purposes, though, I assume that the appropriate goal is neutrality between wages and cash flow. Under the Brady-Ryan plan, neutrality could be achieved by taxing passthrough business cash flow at the same rate as wages under the individual income tax and subjecting all passthrough business cash flow to self-employment tax. The top tax rate on passthrough cash flow would then be just under 36 percent, reflecting 33 percent individual income tax and 2.9 percent Medicare self-employment tax (under the blueprint's assumption that the additional 0.9 percent payroll and self-employment tax introduced by the ACA would be repealed).

It is more difficult to set a comparable level of the corporate cash flow tax rate because corporate cash flow would be subjected to a second layer of tax at the shareholder level. Although dividends paid to top-bracket shareholders would face an effective tax rate of 16.5 percent, the effective tax rate on capital gains resulting from undistributed cash flows would be lower because capital gains taxation would continue to be deferred until realization. If the effective shareholder-level tax rate was around 12 percent, setting the business tax rate at 27 percent would yield a combined effective tax rate around 36 percent, resulting in approximate parity between wages and business cash flow for shareholder-employees.

G. The Inventory Tax

Under current law, inventory costs can be recovered using the first-in first-out or LIFO methods. Although the use of LIFO lowers tax liabilities, section 472(c) imposes a conformity requirement, forbidding taxpayers to use LIFO on their tax returns unless they use it in their financial statements.

Six paragraphs after setting forth the general rule that tangible and intangible assets will be expensed, the blueprint abruptly backtracks on that principle, saying, "With respect to inventory, the Blueprint will preserve the last-in-first-out (LIFO) method of accounting."²⁷ The next sentence reveals unease about the denial of expensing, stating that Ways and Means "will continue to evaluate options for making the treatment of inventory more effective and efficient in the context of this new tax system." No intricate evaluation is required to identify the effective and efficient treatment of inventories under the new system -- inventories should be expensed on the same terms as other capital.

The Brady-Ryan plan's intended treatment of inventories is unclear. The blueprint does not say that the entire inventory accounting system would be preserved, but only that LIFO would be preserved. One possible interpretation is that the conformity requirement would be repealed, making LIFO freely available for all inventories.

The blueprint does not explain why inventories, unlike all other types of capital, would be denied expensing.²⁸ Inventories are often subjected to discriminatory treatment because of a superstition that views them as unproductive.²⁹ In reality, inventories, structures, and equipment are all forms of productive capital, each of which requires an investment and each of which generates a return in the form of increased revenue or cost reductions. Businesses hold inventories because they hold structures and equipment, to make their operations more profitable. Although inventories do not physically produce new goods, they are productive because they allow firms to better meet consumer needs at lower cost by ensuring that items are on hand when consumers order them. The ultimate goal of economic activity is to meet consumer needs, not to increase physical production.

The IRS displayed a partial understanding of the economic function of inventories in a 2010 technical advice memorandum, which concluded that inventories constituted "property held for productive use in a trade or business" for purposes of section 1033(h)(2), which allows tax deferral for property destroyed in a disaster area. Despite an initial assertion that "inventory does not 'produce' property or services within the plain meaning of that term," the memorandum ultimately concluded, after surveying other statutory provisions and legislative history, that inventories are held for "productive use" in the relevant sense.³⁰

For businesses with inventories that grow steadily over time, the use of LIFO results in an effective marginal tax rate equal to the statutory tax rate.³¹ For those businesses, the use of FIFO effectively applies the statutory tax rate to both the real return on inventories and the inflationary appreciation of inventories, resulting in a marginal effective tax rate above the statutory tax rate.³²

Assume that businesses demand a 4 percent real return, net of business-level tax, on investments and that the inflation rate is 2 percent. Under the current tax system, C corporations using LIFO face a 35 percent effective tax rate on inventories. They hold inventories only up to the point at which they yield a real before-tax return (in terms of expense reductions) equal to 6.15 percent, incurring a tax burden of 2.15 percent, 35 percent of the 6.15 percent. C corporations using FIFO because of the conformity requirement hold inventories only up to the point at which they yield a real before-tax return equal to 7.23 percent, incurring a 3.23 percent annual tax burden, 35 percent of the 9.23 percent nominal before-tax return. Because the 3.23 percent annual tax burden is 45 percent of the 7.23 percent real before-tax return, the effective tax rate is 45 percent. The effective tax rate would be even higher at higher inflation rates. Those simple calculations explain why inventories are more heavily taxed than other types of capital under current law.

Although the Brady-Ryan plan would tax inventories more heavily than other types of capital, it would tax them more lightly than under current law. If the blueprint were interpreted to mean that LIFO would be extended to all inventories with no conformity requirement, the effective business-level tax rate on inventories would fall to 20 percent for all C corporations. Corporations would hold inventories up to the point at which they yielded a 5 percent real return, with an annual tax burden of 1 percent (20 percent of the real return).

If the blueprint were interpreted as maintaining the conformity requirement, C corporations using FIFO would hold inventories up to the point at which the before-tax real return was 5.5 percent, with an annual tax burden of 1.5 percent per year (20 percent of the 7.5 percent nominal return). Because the tax burden would be 27 percent of the 5.5 percent before-tax real return, the business-level effective marginal tax rate on inventories would be 27 percent.

The Brady-Ryan plan would therefore reduce the business-level effective marginal tax rate on inventories held by C corporations from a mix of 35 percent and 45 percent to 20 percent or to a mix of 20 percent and 27 percent, depending on how the blueprint is interpreted. (The reduction in effective marginal tax rates would be smaller if the corporate cash flow tax rate was increased to 27 percent, as discussed above.) Despite the improvement from current law, however, there is no justification for imposing an effective tax rate on inventories higher than the zero rate given to all other forms of capital. Astonishingly, the LIFO Coalition, which has been the leading advocate for inventories, professed to welcome the plan's discrimination against inventories, stating, "As long as inventories are required by the tax code, keeping LIFO is the right policy."³³ No doubt, but why should the tax code require inventories under a business cash flow tax?

The plan's inventory tax would distort the allocation of capital by inefficiently discouraging inventory investment. It would also complicate the tax code by maintaining the complexities of inventory accounting and rejecting the elegant simplicity of expensing. Moreover, the inclusion of the inventory tax suggests that tax treatment should vary across different types of capital based on superstitions about the production process. The next step could be to deny expensing to structures based on the superstition that they are unproductive because they just sit there.

H. The Border Adjustment

The plan's border adjustment has received more attention than its other features. In an earlier article, I discussed the economic effects of a border adjustment.³⁴ I briefly review and clarify that analysis and then discuss design issues that the plan needs to address.

1. Simple model.

In a simple model, an immediate permanent uniform border adjustment has no real effects. The simple model assumes perfectly functioning markets, no investments with above-normal returns, and no cross-border asset holdings when the border adjustment is introduced. In the simple model, exports and imports must be equal in present discounted value.

A border adjustment is uniform if the import tax applies at a uniform rate to all imports, the export subsidy applies at a uniform rate to all exports, and the import tax rate is equal to the export subsidy rate when each rate is expressed as a fraction of border prices. The border prices are the net payments crossing international borders, namely the before-tax price of imports and the after-subsidy price of exports. Imports include all sales by foreigners to Americans and exports include all sales by Americans to foreigners.

In the simple model, an immediate permanent uniform border adjustment would immediately and permanently increase the prices (including wages) paid and received by Americans relative to the prices paid and received by foreigners, when those prices are measured in any common currency. Depending on monetary policy, the relative price increase could occur through an increase in the dollar value of prices and wages in the United States, an increase in the foreign currency value of prices and wages in foreign countries, or an increase in the nominal value of the dollar relative to foreign currencies. Because the Federal Reserve would probably not change the U.S. price level in response to a border adjustment and most foreign central banks would probably not change foreign price levels, the most likely outcome would be an appreciation of the dollar. However, the neutrality result would still hold under monetary policies that caused the U.S. price level or foreign price levels, rather than the exchange rate, to change.

In the simple model, the immediate permanent uniform border adjustment would have no effect on trade investment, output, or real wages. The change in Americans' and foreigners' relative prices would offset the boost to exports and the reduction in imports that would otherwise occur. Imports would not become more expensive and exports would not become more lucrative. It is fortunate that the border adjustment would not permanently increase exports and permanently reduce imports. Americans' living standards would be lowered if they forever worked to produce more goods to send abroad for others to enjoy and forever received fewer goods back in return.

The trade neutrality of the immediate permanent uniform border adjustment does not depend upon how the import tax and export subsidy compare with the taxes on domestic products. An immediate permanent uniform export subsidy offsets an immediate permanent uniform import tax imposed at the same rate, regardless of what the rest of the tax system looks like.

The immediate permanent uniform border adjustment would not affect investment in the simple model. Foreigners purchasing U.S. assets would incur higher real costs, but would also receive higher real payoffs. Americans purchasing foreign assets would incur lower real costs, but would also receive lower real payoffs. The impact on costs and payoffs would offset each other for investments with normal returns, which are the only investments that exist in the simple model.

In the simple model, the immediate permanent uniform border adjustment would raise zero revenue in present discounted value because taxed imports are equal to subsidized exports in present discounted value. Because the border adjustment would not alter the economy's equilibrium, it would not harm or help anyone and therefore would not impose taxes on, or make transfer payments to, any persons. To be sure, the border adjustment would raise revenue in years with trade deficits and reduce revenue in years with trade surpluses. Those revenue effects would arise because, in line with the effects described in the preceding paragraph, the government would effectively borrow from foreigners as they bought U.S. assets (repaying them as they received payoffs on the assets) and would effectively lend to Americans as they bought foreign assets (collecting repayment from them as they received payoffs on the assets).

In the simple model, an immediate uniform permanent border adjustment would have no real effects. There would therefore be no reason to support, and no reason to oppose, that policy.

The use of the simple model is not intended to suggest that a border adjustment would actually have no real effects. Instead, the simple model serves two purposes. First, it explains why the border adjustment, contrary to the claims of some of its supporters, would not permanently boost exports and permanently reduce imports. Second, it reveals that a border adjustment's real effects must be because of its impact on above-normal-return investments, initial cross-border asset holdings, or market imperfections (none of which were present in the simple model) or must be attributable to a lack of immediacy, permanence, or uniformity.

2. General model.

In a general model that includes above-normal returns and initial cross-border asset holdings, trade need not balance in present discounted value. Instead, the present discounted value of U.S. trade deficits equals the value of Americans' initial holdings of foreign assets plus the present discounted value of Americans' future above-normal returns from foreign investments minus the value of foreigners' initial holdings of U.S. assets minus the present discounted value of foreigners' future above-normal returns from U.S. investments.

In the general model, an immediate permanent uniform border adjustment could have positive or negative revenue effects, in present discounted value, depending on the values of initial cross-border asset holdings and cross-border above-normal returns. Foreigners' holdings of U.S. assets exceed Americans' holdings of U.S. assets, indicating that the border adjustment would likely lose revenue in present discounted value. However, part of that revenue loss would be offset by the tendency of Americans to earn (or report) greater above-normal returns abroad than foreigners earn (or report) in the United States.

In computing the revenue effect of adding a border adjustment to a non-border-adjusted business cash flow tax, the relevant above-normal returns to use in the calculations are those that would be reported under the non-border-adjusted tax, regardless of where the returns were really earned. The border-adjusted cash flow tax would collect tax on all of Americans' above-normal returns while the non-border-adjusted cash flow tax would collect tax only on the above-normal returns booked in the United States. The border adjustment therefore would raise revenue from the difference between the two numbers, even if the latter number were artificially low because of income shifting.³⁵

The border adjustment would impose a tax on above-normal returns that Americans earned on foreign assets. It would also impose a tax on foreign assets that Americans held when the border adjustment was introduced. As previously described, the border adjustment would reduce Americans' real payoffs on foreign assets. That reduction would not be fully offset by a reduction in the assets' real costs, either because the costs are smaller than the payoffs in present discounted value (for assets with above-normal returns) or because the costs were incurred before the border adjustment was introduced (for initial holdings).

Conversely, the border adjustment would provide a subsidy to above-normal returns that foreigners earned on U.S. assets. It would also provide a subsidy to U.S. assets that foreigners held when the border adjustment was introduced. As previously described, the border adjustment would increase foreigners' real payoffs on U.S. assets. That increase would not be fully offset by an increase in the assets' real costs, either because the costs are smaller than the payoffs in present discounted value (for assets with above-normal returns) or because the costs were incurred before the border adjustment was introduced (for initial holdings).

To a first approximation, the interaction of a border adjustment and a business cash flow tax can be described as follows. A non-border-adjusted business cash flow tax would be origin-based and would therefore impose tax on above-normal returns on U.S. assets earned by both Americans and foreigners and on both groups' initial holdings of U.S. assets. Applying a uniform border adjustment, with the same import tax rate and export subsidy rate as the cash flow tax rate, would add Americans' above-normal foreign returns and initial foreign assets to the tax base and would remove foreigners' above-normal U.S. returns and initial U.S. assets from the tax base. The border-adjusted cash flow tax would therefore impose a tax on all of Americans' above-normal returns and initial assets, making it destination-based.

The above description is imprecise, however, because it assumes that the incidence of a tax on initial assets falls on those who will consume from those assets. In general, however, the incidence depends on the contractual terms of the assets, which were negotiated before the tax took effect. If the contractual terms depend on prices or nominal exchange rates, monetary policy may affect the incidence. For example, as discussed above, equity holders are likely to bear the entire transitional burden of a cash flow tax because debt holders have a contractual right to fixed payments.

Consider the border adjustment's revenue collection from Americans' initial holdings of foreign assets and maintain the standard monetary policy assumption (the Federal Reserve and foreign central banks keep price levels unchanged, so that the dollar appreciates). Then, if the foreign assets had fixed nominal values denominated in foreign currency, the American holders would bear the tax burden through a reduction in real asset values. However, if the foreign assets had fixed nominal values denominated in U.S. dollars (perhaps because they were issued by developing countries that found it necessary to borrow in dollars rather than in their own currencies), the foreign issuers would bear the tax burden through an increase in the real value of their liabilities. Similarly, consider the border adjustment's revenue loss from foreigners' initial holdings of U.S. assets. Under the standard monetary policy assumption, the benefit of that revenue loss would be captured by the foreign holders through increased asset values if the U.S. assets had fixed nominal values denominated in U.S. dollars, but would be captured by the American issuers if the U.S. assets had fixed nominal values denominated in foreign currencies.³⁶ Other outcomes would occur for different types of assets (such as equity) or under different monetary policy assumptions.

The border-adjusted cash flow tax, unlike the non-border-adjusted cash flow tax, would not impose a tax penalty on making above-normal-return investments in the United States rather than abroad. Americans would pay U.S. tax on above-normal returns no matter where they were earned and foreigners would be exempt from U.S. tax on above-normal returns no matter where they were earned. Similarly, the border-adjusted cash flow tax would not provide a tax incentive to book above-normal returns abroad through income shifting schemes.

In the general model, the border adjustment would reduce the present discounted value of trade deficits by increasing foreigners' above-normal-return investments in the United States and reducing Americans' above-normal-return investments abroad. Although the change in investment flows would initially increase the trade deficit, the subsequent payoffs on the investments would reduce the trade deficit. Because the present discounted value of the payoffs on above-normal-return investments exceeds the initial investment, the present value of trade deficits would fall. The border adjustment would also reduce the measured trade deficit by encouraging above-normal returns to be booked in the United States rather than abroad.

3. Anticipated, temporary, or nonuniform.

A border adjustment would have important real effects if it did not take effect immediately after announcement, if it was expected to be temporary, or if it did not apply uniformly to all imports and exports.

A border adjustment that was expected to be temporary would reduce the trade deficit while it was in effect and would cause a slight ongoing increase in the trade deficit after it ended. A border adjustment that applied to some products, but not others, would reduce imports and increase exports for the covered products, but would increase imports and reduce exports for other products. A border adjustment that imposed an import tax rate higher than the export subsidy rate would be equivalent to an import tariff.

After a border adjustment (or the possibility thereof) was announced and before it took effect, Americans would increase imports to avoid the looming import tax and would reduce exports to await the impending export subsidy, causing the trade deficit to rise during that interval. The increased trade deficit would be financed by larger capital inflows, through an increase in Americans' consumption and investment in the United States.³⁷ The increased consumption and investment would reduce the effective-date value of American-owned foreign assets, which would be taxed by the border adjustment, and would increase the effective-date value of foreign-owned U.S. assets, which would be subsidized by the border adjustment. However, there would be no tax incentive for Americans and foreigners to swap existing assets before the border adjustment took effect because such swaps could not increase foreign-owned U.S. assets without increasing American-owned foreign assets.

Because of the increase in foreign-owned U.S. assets and reduction in American-owned foreign assets, there would be a slight ongoing reduction in the trade deficit after the border adjustment took effect, which would offset, in present discounted value, the higher trade deficits before implementation. The dollar would begin to strengthen, or the relative prices paid and received by Americans would otherwise begin to rise, when the potential border adjustment was announced.

4. Evolution of the debate.

In past years, supporters of border adjustment primarily argued that it would permanently increase exports and reduce imports. When economists pointed out that those alleged trade effects would be negated by an increase in the relative prices paid and received by Americans (probably manifested through an appreciation of the dollar), they were viewed as opposing the border adjustment. During the debate over the Brady-Ryan plan, however, many supporters of border adjustment have acknowledged the relative price changes and even cited them when (correctly) arguing that importers would not be harmed by the border adjustment. Yet, many of those supporters continue to argue that the border adjustment would level the playing field for U.S. products in the face of foreign VATs, an argument that is invalid because it fails to recognize the change in relative prices. Opponents are now more likely to argue that the relative price changes would not occur.

In some ways the debate seems to be moving closer to economic reality. More attention is being paid to the border adjustment's effects on income shifting, the location of above-normal-return investments, and the treatment of initial cross-border asset holdings.

5. VAT expertise.

The border adjustment would face many challenging design issues. For example, it would be necessary to determine whether purchasers were Americans or foreigners, which could be difficult for purchasers of some services, and the appropriate tax treatment of mergers and asset acquisitions must be determined. In most cases, however, it would not be necessary for the United States to reinvent the wheel. More than 160 other countries have border-adjusted VATs, many of which have been in place for decades, and they have found ways to resolve those issues.

To be sure, those countries' VATs use the credit-invoice method and do not provide a wage deduction. For many issues, though, those differences are of limited importance. The danger is that Republicans might fail to take advantage of the VAT expertise out of an unwillingness to admit their plan's similarity to a VAT and needlessly seek to reinvent the wheel.

6. Departures from uniformity.

A highly simplified representation of the Brady-Ryan plan's border adjustment treats it as imposing a tax at a 20 percent tax-inclusive rate on imports and providing a subsidy at a 20 percent subsidy-exclusive rate for exports. (The representation assumes that all imports and exports are done by C corporations.) The import tax and the export subsidy would then be 25 percent of the border prices. An immediate permanent border adjustment with this uniform rate would increase the prices paid and received by Americans by 25 percent relative to the prices paid and received by foreigners. If the relative price change occurred through a change in the exchange rate, the dollar would strengthen by 25 percent against foreign currencies.

In practice, numerous factors could prevent the plan's border adjustment from being completely uniform. For example, some imports are purchased directly by American consumers and would need to be subjected to a separate tax, which might be difficult to enforce.³⁸ A separate tax would need to be imposed on imports purchased by tax-exempt entities, including state and local governments and nonprofit organizations.

Like VATs around the world, the import tax could not be applied to all purchases that Americans made from foreigners and the export subsidy could not be provided to all sales that Americans make to foreigners. As my American Enterprise Institute colleague Stan Veuger has noted, the export subsidy would probably not be provided to foreign tourists' purchases in the United States and foreign students' tuition payments to U.S. educational institutions.³⁹ The border adjustment would discourage those exports because they would face the headwinds of a stronger dollar without any offsetting subsidy.

Political pressure could lead to other departures from uniformity. Proposals to exempt specific industries from the border adjustment, or to allow businesses to elect whether to be subject to the border adjustment, must be resisted at all costs.

Because C corporations would face a 20 percent tax rate and passthrough businesses would face a maximum 25 percent rate, the plan would create an incentive for C corporations to engage in imports and for passthrough businesses to engage in exports. Uniformity could be attained by applying the border adjustment at a consistent 20 percent rate. Passthrough businesses would receive a 20 percent subsidy on exports, which would be less generous than an exemption of export sales, and would pay a 20 percent tax on imports, which would be less onerous than a denial of deduction for import costs. If the tax rate on passthrough cash flow was raised to 33 percent and the tax rate on corporate cash flow was raised to 27 percent, as suggested above, the border adjustment could be applied at a uniform 27 percent rate. An immediate permanent uniform border adjustment applied at that rate would cause the prices paid and received by Americans to rise 37 percent relative to the prices paid and received by foreigners.

A larger potential departure from uniformity concerns the treatment of exporters' negative tax liabilities. Businesses with significant exports could easily have negative tax liabilities. Suppose that a business paid $50 of wages, bought $30 of items from other domestic businesses, and made $100 of sales. The business would have a negative tax liability if more than $20 of its sales were exports.

If the negative tax liability were refunded in cash, the exporter would face a 20 percent subsidy rate, which would match the 20 percent import tax rate. In the absence of cash refunds, however, the effective export subsidy rate would be smaller than the tax rate on imports, which would cause the border adjustment to impede trade. The excess of the import tax rate over the effective export subsidy rate would have the same economic effects as a simple import tariff. Also, if the absence of cash refunds affected some industries more severely than other industries, additional departures from uniformity, and additional distortions, would be introduced because some exported products would be favored over other exported products.

Because many exporters would have negative tax liabilities in perpetuity, the problem could not be corrected by carrybacks (which the plan would not offer) or carryforwards (even carryforwards with interest, as provided in the plan). A recent study by Elena Patel and John McClelland of the Treasury Department used data on a large sample of businesses to examine the magnitude of the problem.⁴⁰ They estimated that a border-adjusted cash flow tax would have had a $13.08 trillion cumulative base for 2004 through 2013 if no tax relief were provided for negative annual tax liabilities. With full refunds, the cumulative tax base would fall to $9.44 trillion. Allowing losses to be carried back and forward without limit within the 2004-2013 period would shrink the tax base to only $12.5 trillion. In other words, carrybacks and carryforwards would allow only $580 billion of the $3.64 trillion of negative cash flows to be used. A large portion of the $3.06 trillion of unused negative cash flows would be because of the border adjustment. Patel and McClelland found that a non-border-adjusted cash flow tax would have only $1.76 trillion of unused negative cash flows and that a stylized version of a business income tax would have only $1.5 trillion of unused losses.

Without refunds or a similar policy, exporters would be driven to merge with other businesses, whose positive tax liabilities could be used to offset negative tax liabilities arising after the merger. (Unless section 382 was amended, the merger would offer only limited ability to offset carryforwards of negative tax liabilities arising before the merger.) Importers would presumably be the most attractive merger partners because they would have large tax remittances. Such tax-motivated mergers would be costly, would take time to arrange, and would undermine efficient business allocation. Moreover, many exporters might fail to find merger partners and would not be able to fully monetize their negative tax liabilities. Other possible arrangements include having importers or other companies serve as export brokers. Some export opportunities might also shift from businesses with negative tax liabilities to those with positive tax liabilities. All of those responses would be limited and inefficient. Moreover, legislative or regulatory steps might be taken to curb some of those responses as abusive tax practices.⁴¹

For the border adjustment to function properly, exporters must be able to monetize their negative tax liabilities. Under the Growth and Investment Tax plan proposed by the tax reform advisory panel in 2005, exporters would have received refunds for their negative tax liabilities, although other businesses with negative cash flows would have received unlimited carryforwards with interest.⁴² Patel and McClelland noted that refunds are the "theoretically appropriate" response, but cautioned that non-legitimate businesses might seek relief for phantom losses. They noted that refunds could be allowed against businesses' payroll tax liabilities or their tax withholding payments, thereby ensuring that money is paid only to claimants with actual business operations.⁴³

Cash refunds might also encounter political obstacles if they were perceived as corporate welfare. In reality, cash refunds would merely compensate for the reduction in exporters' sale proceeds caused by the stronger dollar. Export refunds are routinely provided under VATs without controversy. The provision of relief for exporters with negative tax liabilities should be a high priority in the design of the border adjustment.

7. Transitory revenue gain.

Because the United States is running a trade deficit, the border adjustment would yield a short-term revenue gain. The Tax Foundation estimated that the border adjustment would raise $1.07 trillion during fiscal 2016 through 2025 on a static basis.⁴⁴ The TPC estimated that the border adjustment would raise $1.18 trillion during fiscal 2017 through 2026 and another $1.69 trillion during fiscal 2027 through 2036.⁴⁵ The net revenue gain within the 10-year budget window appears to account for much of the political support for border adjustment.⁴⁶

As discussed above, however, the revenue gain is transitory. In the simple model, an immediate permanent uniform border adjustment would raise no revenue in present discounted value. Patel and McClelland aptly commented that "trade flows must eventually balance. If the United State is currently a net importer, in the longer run it would be a net exporter. This implies that at some point in the future, the border adjustments reduce the tax base rather than increase it."⁴⁷

The transitory revenue gain is not an unambiguous political advantage because some critics mistakenly see the revenue as proof that the border adjustment would impose a tax increase on American businesses that buy imported goods and their American customers.⁴⁸ Some supporters mistakenly see the revenue as proof that the border adjustment would impose a tax increase on foreign producers of items sold in the United States.⁴⁹ All of these arguments are invalid because the border adjustment would not change either the real prices that Americans paid for imports or the real prices that foreigners received for selling them.

As explained above, in the simple model, the government's financial inflows and outflows reflect borrowing and lending on cross-border investments. Under the simple model's assumption, the transitory revenue gain should be excluded from distributional analysis, just as the government's proceeds from borrowing are excluded.⁵⁰

Although a theoretical case could be made for also excluding the transitory revenue gain from the revenue estimate, that practice would be incompatible with the treatment of other forms of disguised borrowing and lending.⁵¹ For example, expensing imposes zero net tax on a marginal business investment and raises zero net revenue, but features an initial outflow of government funds followed by subsequent inflows of funds. It is equivalent to the government lending to the business when the investment is made and collecting repayments as the investment's payoffs occur. Yet, standard practice includes the initial outflow as a revenue loss rather than as lending. Symmetry requires inclusion of the temporary revenue gains from border adjustment, particularly when analyzing a plan that features both expensing and border adjustment.

8. Trade policy implications.

President Trump has called for tariffs, which would undermine global free trade and harm both the United States and its trading partners. Some congressional Republicans have expressed the hope that the border adjustment could serve as a substitute for the tariffs advocated by Trump.⁵² During a January 4 radio interview, Ryan argued that tax reform that leveled the playing field (an apparent reference to border adjustments) would be preferable to tariffs.⁵³ House Majority Leader Kevin McCarthy, R-Calif., has similarly argued that the Brady-Ryan tax plan as a superior alternative to tariffs.⁵⁴

Marketing a policy on specious grounds generally leads to bad outcomes. For example, Republicans who marketed high-income marginal tax rate reductions as a special benefit to small businesses laid the groundwork for proposals that offered distortionary subsidies to small businesses.⁵⁵ Marketing the border adjustment, to Trump or to any other audience, as a protectionist measure would only strengthen protectionist pressures by reinforcing the widespread misperception that trade deficits are inherently undesirable. If that premise were accepted, it would leave no basis on which to argue against tariffs.

Marketing the border adjustment as protectionist would also make it harder to defend before the World Trade Organization. Brady and Ways and Means Tax Policy Subcommittee Chair Peter J. Roskam, R-Ill., recently expressed confidence that the border tax adjustment would be found to comply with WTO rules,⁵⁶ echoing a view expressed in the blueprint.⁵⁷ That view has won little support, however. The TPC observes that many commentators believe that the border adjustment would be found to violate WTO rules.⁵⁸ The 2005 tax reform advisory panel recognized the uncertainty over whether the border adjustment would be allowed under international trade rules.⁵⁹ Michael Graetz also argued that the panel's proposal would be inconsistent with international trade agreements.⁶⁰

Gary Hufbauer and Zhiyao Lu recently provided a thorough analysis of the WTO rules.⁶¹ Separate legal regimes apply to imports and exports and to goods and services. Hufbauer and Lu conclude that the tax on imported goods and the subsidy to exported goods would clearly violate WTO rules, but that the tax on imported services might be permissible and the subsidy to exported services would not run afoul of any restrictions.

Article III of the General Agreement on Trade and Tariffs provides that imported goods may be subjected to any tax that is imposed, to the same extent, on domestic products. Hufbauer and Lu note that it is unclear whether the WTO would consider the cash flow tax to be a tax on products. Although the WTO might view the cash flow tax as not being a tax on products, which would categorically preclude any border adjustment, it might consider it sufficiently similar to a VAT to constitute a tax on products. Nevertheless, Hufbauer and Lu conclude that the tax on imported goods in the Brady-Ryan plan would still be invalid because imported goods would be taxed on their full market value while domestic products would be taxed only on their market value minus wage costs. Article XVII of the General Agreement on Trade in Services requires that imported services must receive "treatment no less favorable" than the treatment of domestic services. Hufbauer and Lu note that the reference to treatment is more general than the specific requirement that GATT imposes on goods. They conclude that the Brady-Ryan plan's tax on imported services may be permissible because the total taxes imposed on the production of domestic services, including payroll taxes, may be similar to the tax on imported services.

Annex I to the Agreement on Subsidies and Countervailing Measures provides that direct taxes, but not indirect taxes, may be rebated when goods are exported. Hufbauer and Lu note that the cash flow tax may be classified as an indirect tax, but that the export subsidy would still be prohibited because the subsidy exceeds the amount of cash flow tax imposed on the product. Hufbauer and Lu note that no binding restrictions on subsidies for exported services have yet been adopted.

The validity of the subsidy to exported services and the possible validity of the tax on imported services offers little comfort. Because the neutrality of the border adjustment requires that it be uniform, the application of the border adjustment only to services is not a viable option.

It would similarly be undesirable to impose an import tax and provide an export subsidy that approximates the tax on domestic products. The import tax and export subsidy would no longer be uniform across different products, destroying its economic neutrality.

The WTO rules lack a solid economic foundation. As I explained in my previous article, a uniform import tax accompanied by a uniform export subsidy at the same rate does not distort trade. Countries should therefore be allowed to adopt such provisions, regardless of what domestic tax system, if any, they may have. If the border adjustment is permitted for a VAT, it should be permitted for a VAT with a wage deduction.

Nevertheless, the United States must deal with the WTO rules that exist, not those that would have been adopted in a more perfect world. The European Union and other U.S. trading partners recently began laying the groundwork for a WTO challenge to the border adjustment. Chad Brown of the Peterson Institute estimates that an unfavorable ruling could lead to $385 billion of annual trade retaliation against the United States, far surpassing the scope of the retaliation authorized in any previous WTO case.⁶²

9. Should it stay or should it go?

A properly designed border adjustment offers significant advantages by eliminating income shifting and attracting above-normal-return investments to the United States. Nevertheless, the border adjustment has a severe disadvantage. As discussed above and in my earlier article, it grants a large tax cut, at the expense of the U.S. Treasury, to foreigners holding U.S. assets when the border adjustment is introduced. The wealth effects have started to receive more attention in the debate.⁶³

Moreover, the difficulty of providing exporter refunds suggests that the border adjustment might not be properly implemented. The prospect of WTO sanctions is another disadvantage. Moreover, rushing through the border adjustment during a short time period may complicate getting the design issues right, particularly if Congress fails to tap the expertise that other countries have developed with VATs.

An alternative path would be to adopt a non-border-adjusted cash flow tax, which would retain many of the advantages of the border-adjusted cash flow tax. The debt-equity distortion would still be largely eliminated, the tax penalty on saving would still be reduced, and the tax penalty on locating normal-return investments in the United States would still be eliminated through expensing. The loss of the border adjustment's transitory revenue gain could be offset by higher tax rates on business cash flow. Under the non-border-adjusted tax, however, there would be a tax penalty on locating above-normal-return investments in the United States and a tax incentive for income shifting.

The choice of whether to border adjust poses difficult trade-offs. In 2012 Carroll and I recommended against border adjusting the business cash flow tax under an X tax, in order to avert the windfall tax cut for foreigners holding U.S. assets.⁶⁴ Our analysis may have given too little weight to the border adjustment's effects on above-normal-return investments and income shifting and thereby understated its advantages. Nevertheless, the recent debate has revealed significant political and implementation challenges that weigh against the border adjustment. On balance, the case for the border adjustment remains uncertain, at best. At a minimum, it is clear that the overall tax reform effort should not be held hostage to the border adjustment.

10. The VAT alternative.

If the border adjustment is retained, it might be desirable to relabel the cash flow tax as a VAT. Although the blueprint boasts that it "does not include a value-added tax,"⁶⁵ the plan can easily be rewritten to include a VAT, as I discussed in my earlier article.

As explained above, an X tax is equivalent to a VAT, modified to attain progressivity through separate taxation of wages and business cash flow. A consumption tax that completely replaces the income tax system should be designed as an X tax rather than a VAT to maintain the tax system's progressivity. But, that arrangement is unnecessary if the plan retains a significant income tax component, which can be used to attain progressivity. The House plan keeps a tax, with a 16.5 percent top rate, on interest, dividends, and capital gains.

The VAT alternative would have several advantages. First, the border adjustment would clearly be permissible under WTO rules. Second, the tax could be implemented using the credit-invoice method, which would yield enforcement advantages. Third, the perceived problems associated with tax increases on importers and windfalls for exporters would disappear because the pattern of tax remittances under a VAT appears to arouse little or no political concern. Fourth, the price increase would eliminate any need for the nominal exchange rate to adjust, perhaps alleviating concern about whether and when the exchange rate would respond. Fifth, the increased price level would spread the transitional burden evenly over equity holders and debt holders, which might distribute the burden more fairly among Americans and which would limit the giveaway to foreigners, who are more likely to hold debt.⁶⁶ It is possible, however, that the relabeling could spur the growth of government spending because it would combine cash flow taxation and wage taxation into a single large VAT.

I. Conclusion

The House plan could promote economic growth by partially moving the tax system toward a Bradford X tax, but some changes should be made. Raising the tax rates on business cash flow would help eliminate any revenue shortfall and maintain progressivity. The plan's discriminatory inventory tax should be eliminated. Making modest changes would allow the plan to fulfill its growth potential.